Bittensor Surges as NVIDIA CEO Highlights Decentralized AI Training

Bittensor's TAO token has experienced a dramatic surge in value after Jensen Huang, CEO of NVIDIA, publicly acknowledged the project's potential during an episode of the All-In Podcast. Following this endorsement, TAO saw a 17% increase in a single trading session, reaching intraday highs of over $300, its highest level since January. Over the past week, the token's value has appreciated by 37%, and it has climbed 55% in the last 30 days.

The Covenant-72B Moment

To understand the significance of this movement, consider the implications of Covenant-72B, a large language model fully trained using decentralized infrastructure. Unlike traditional models that rely on centralized data centers and substantial budgets, Covenant-72B was developed collaboratively by over 70 independent contributors using standard internet hardware. This innovative approach enabled the successful training of a 72-billion-parameter model on 1.1 trillion tokens, achieving a notable 67.1 MMLU score, as documented in a March 2026 arXiv paper.

What distinguishes Covenant-72B is its performance. With a 67.1 MMLU score, it competes with Meta's Llama 2 70B and early GPT-4-class models on standardized tests, all without a $100 million data center budget. This validation of decentralized training methods has garnered the attention of key players in the AI field.

Market Dynamics and Future Outlook

Despite these impressive gains, market indicators present a more complex picture. TAO recently closed at $272.9, dipping below the DEMA 9 level of $278.9. Current sentiment on the order book shows a slight leaning towards sellers, with a 41/59 ratio. While the MACD indicator remains positive, it is showing signs of fading momentum. The immediate market reaction to Huang's comments reflects strong interest, yet questions remain about the sustainability of this price surge in the long run.

For investors and holders of TAO, Huang's remarks provide significant validation of Bittensor's approach to decentralized AI training. The event can be compared to the SEC's approval of a Bitcoin ETF; while it does not alter the underlying technology, it shifts the narrative and increases visibility. As the AI sector continues to evolve, the fundamental question remains: will the current excitement around decentralized models lead to lasting value, or will market optimism fade as investors reassess the fundamentals driving this technology?